Vol. 24, No.2 Printed in U.S.A.
JOURNAL OF VIROLOGY, Nov. 1977, p. 590-601 Copyright © 1977 American Society for Microbiology
Changes in the Capsid Structure and Stability of Defective Particles of Bacteriophage R17 WALLACE J. IGLEWSKI Department of Microbiology and Immunology, University of Oregon Health Sciences Center, Portland, Oregon 97201 Received for publication 28 April 1977
Serological and chemical methods were used to compare the capsid structure and stability of R17 phage and amA31 defective particles. Immunodiffusion analysis demonstrated identity between intact R17 and amA31 capsids and between dissociated subunits of both R17 and amA31 and purified coat protein. Radioimmunoassays detected an antibody in R17 antisera that binds to intact R17 but could not be absorbed from R17 antisera with amA31. The R17 antibody remaining in amA31-absorbed sera did not neutralize infectivity of R17 phage. Differences between the surface composition of R17 and amA31 capsids were also detected by iodination. Capsids of R17 bound approximately four times more 125I than amA31, which was accounted for by a decreased 125I labeling of coat protein. Finally, amA31 capsids dissociated under milder conditions of sodium dodecyl sulfate treatment than R17 capsids. The sodium dodecyl sulfate dissociation of both R17 and amA31 capsids resulted in the formation of a transient 38,000-dalton intermediate, which subsequently dissociated to coat protein monomers. Preparations of dissociated R17 capsids also contained assembly protein and a protein of approximately 27,000 daltons. The 27,000-dalton protein was also found in preparations of dissociated amA31 capsids.
Bacteriophage R17 is a small icosahedral phage having a diameter of about 13.5 nm (9). The phage has a single-stranded RNA genome, which is surrounded by 180 coat protein subunits and 1 assembly protein subunit for attachment to susceptible cells (7, 14, 21, 25). Recently, it has been shown that assembly protein is located on the surface of the phage (2), which is compatible with its role as an attachment site. In terms of the integrity of the phage, assembly protein participates in maintaining the barrier between the phage genome and the external environment, since the genome of mutants that lack assembly protein is degraded by treatment with RNase (8, 14). In this study the organization and stability of the coat protein capsid of the R17 phage was compared with an amber mutant that lacks assembly protein. The absence of assembly protein alters the surface of the capsid and decreases the stability of the capsid as measured by dissociation with sodium dodecyl sulfate (SDS). Thus, assembly protein appears to also have a general effect on the capsid structure. MATERIALS AND METHODS
richia coli S26 and S26RIE, respectively, were provided by R. Kamen. The viral mutant was grown and plaqued on E. coli S26RIE by the method of Tooze and Weber (24) and contained less than 1.0% revertants. E. coli S26 was used to grow wild-type R17 and produce defective particles of amA31 lacking the assembly protein after infection with complete amA31 grown in the permissive cells. Growth and purification of wild-type R17 and defective amA31 phage. Growth of large batches of R17 and its purification are based on the published procedure of Vasquez et al. (25). Briefly, batches (20liter) of E. coli S26 grown in MS broth (20) to 3 x 108 cells per ml were infected with 10 PFU of R17 per cell. After 3 h of incubation, the cells were chilled and lysed by shaking with 10% (vol/vol) chloroform. The lysate was decanted, and 350 g per liter of ammonium sulfate was added. The precipitate was harvested by continuous-flow centrifugation in a Sorvall SS34 rotor at 18,000 rpm. The pellet was resuspended in 300 ml of 0.1 M NaCl and 0.05 M Tris, pH 7.2 (ST), buffer, and the phage was further purified by two cycles of differential centrifugation at 10,000 rpm in a Sorvall SS34 rotor for 20 min to remove debris and at 50,000 rpm in a Beckman SW50 rotor for 90 min to sediment the phage. The phage was resuspended in ST buffer containing 10-3 M MgCl2 and treated with a mixture of 5,ug of RNase A and 5 ,ug of DNase per ml for 30 min at 37°C. This was followed by digestion with 25 ,g of Pronase per ml at 25°C for 30 min. The phage was sedimented as described above and resuspended in ST buffer for CsCl equilibrium
Bacteria and bacteriophage. Amber mutant amA31 of the bacteriophage R17, along with the nonpermissive and permissive indicator strains, Esche590
VOL. 24, 1977
CAPSID STRUCTURE OF R17 DEFECTIVE PARTICLES
centrifugation at 39,000 rpm in a Beckman SW50 rotor for 24 h. The virus that banded at p = 1.43 was collected and dialyzed against ST buffer before being centrifuged through a 5 to 20% sucrose gradient containing ST buffer (11) at 27,000 rpm for 3.5 h at 5VC in a Beckman SW27 rotor. Phage bands in both CsCl and sucrose gradients were monitored by optical density at 280 nm (OD20) and PFU per fraction. The buoyant density of the phage was determined by measuring the CsCl concentration of fractions with a refractometer. Defective amA31 phage particles were produced in E. coli S26, the nonpermissive host, and purified as described above. In the CsCl gradient, the defective particles banded at a buoyant density slightly lower than infectious units due to a partial loss of their RNA genome during treatment with RNase A (21). Likewise, the defective particles sedimented slightly slower than the back-revertant, wild-type phage in a sucrose gradient (9). Selection of defective-particle fractions with minimum infectivity per unit of adsorbance at 280 nm within the phage band resulted in partial purification of defective particles from intact phage. The final step in purification of the amA31 defective particles from wild-type phage made use of the inability of assembly protein-deficient mutants to adsorb to E. coli S26 (14). Cells were grown to 4 x 108 cells per ml in MS and harvested by centrifugation at 10,000 rpm for 10 min in a Sorvall SS34 rotor. The cells were washed and resuspended in physiological saline containing 5 ml of 1 M CaCl2 and 2 ml of 50% glucose per liter. The cells were collected on a 142mm membrane filter (Millipore Corp.), 0.22-,um pore size. Eight sheets of Whatman no. 4 filter paper were used as a prefilter to prevent clogging of the Millipore filter. Glucose and CaCl2 were added to the phage suspension at the concentrations indicated above, warmed at 37°C, and slowly filtered through the layer of cells. The defective particles passed through the filter, whereas the revertant, wild-type R17 adsorbed to the cells as expected (14). The cell layer was washed with physiological saline containing CaCl2 and glucose to flush through residual defective phage. Purity of amA31 preparations was determined by the decreased number of PFU per OD280 of the phage suspension and SDS-polyacrylamide gel electrophoresis of nIlabeled protein from SDS-disrupted phage. Lactoperoxidase-catalyzed iodination of R17 and amA31 defective phage. Enzymatic iodination of purified phage particles was done essentially as described by Gibson (6). To 100,ul of purified phage containing 100 Mg of protein, the following sequential additions were made: (i) 10 IlI of carrier-free '5I; (ii) 20 M1 of lactoperoxidase bound to Sepharose beads (approximately 10,ug of bound lactoperoxidase); (iii) 10Ml of 3.3 x 10-5 M KI and 10 ,ul of a 1:2,000 dilution of 30% H202 in distilled water. The reaction was allowed to proceed for 30 min at 22°C with gentle shaking to keep the Sepharose beads suspended. Under these conditions the reaction should be essentially completed in 10 to 15 min (3). The Sepharose beadbound lactoperoxidase was removed by centrifugation at 500 x g for 10 min at 5°C. Free "1I was removed during subsequent rate-zonal sucrose gradient sedimentation of the phage as described above. The phage
591
band was diluted in ST buffer and then pelleted by centrifugation at 50,000 rpm for 90 min in a Beckman SW50 rotor. Preparation of R17 coat protein. The cold acetic acid method of Sugiyama et al. (22, 23) was used to isolate coat protein from purified R17 phage. One volume of ice-cold R17 phage (6 mg/ml) in ST buffer was added to 2 volumes of cold glacial acetic acid and incubated in an ice bath for 1 h. The precipitated RNA and assembly protein were then removed by centrifugation at 8,000 rpm in a Sorvall SS34 rotor for 20 min at 5°C. The protein solution was dialyzed against three to four changes of 0.001 M acetic acid at 5°C until the pH of the dialysate reached 3.2 to 3.5. The coat protein solution was centrifuged at 50,000 rpm in an SW50 rotor at 5°C for 90 min to remove possible traces of phage particles that escaped the acid treatment and then dialyzed extensively against a buffer containing 0.01 M sodium phosphate (pH 7.2), 0.29 M /?-mercaptoethanol, and 0.1% SDS. The coat protein was stable for several weeks at room temperature when stored under nitrogen in a sealed tube. Preparations of l"I-labeled coat protein were obtained from previously iodinated phage. SDS-polyacrylamide gel electrophoresis. Purity of phage preparations was examined by analysis on SDS-polyacrylamide gels. The 17- by 0.8-cm or 8by 0.8-cm gels were prepared and run according to the method of Weber and Osborn (27), except the reducing agent was omitted when phage protein was analyzed. Phage samples were dissociated by heating for 1 min at 90°C in 1.0% SDS. The protein bands in the gels were fixed and stained in 0.7% Coomassie brilliant blue, and the gel was destined by diffusion in a solution of 20% methanol, 10% acetic acid, and 70% water. Proteins labeled with "1I were detected by counting gel fractions in a Beckman Biogamma counter. Purification of assembly protein. Assembly protein was purified according to the procedure of Steitz (21). The procedure consists of solubilizing purified phage in 6 M guanidine-HCl, dialyzing into 8 M urea, and eluting from a cellulose phosphate column with a linear NaCl gradient. Assembly protein elutes in 8 M urea-0.05 M Tris-hydrochloride (pH 8.5) containing 0.2 M NaCl. The protein must be kept in high concentrations of urea, guanidine-HCl, or SDS to prevent precipitation. Preparation of antibodies. Antisera to R17 and amA31 were made in 2-kg New Zealand white rabbits by injecting 0.1 mg of phage in complete Freund adjuvant into four subcutaneous sites on the back and in two rear footpads. The rabbits were immunized in the same manner 3 weeks later with phage in incomplete Freund adjuvant and bled after 10 days. Antiserum to assembly protein was made by emulsifying the assembly protein fractions (0.1 mg of protein) from the cellulose phosphate column with an equal volume of complete Freund adjuvant and inoculating and bleeding rabbits on the same schedule as described above. Immunodiffusion analysis. Immunodiffusion (15) was done at room temperature on glass microscope slides using 0.75% agarose in 0.01 M phosphate
592
IGLEWSKI
buffer (pH 7.5) containing 0.5 M glycine, 0.14 M sodium citrate, and 0.02% Merthiolate. Approximately 2.5 ml of the agarose solution (sometimes containing 0.1% SDS) was layered onto each slide and allowed to harden, and wells were cut out of the agar. Incorporation of SDS in diffusion agar helped prevent aggregation and precipitation of coat protein antigen when SDS-dissociated phage antigens were used (28). The slides were placed in a moist chamber to retard drying. Radioimmunoassay for phage antibody. Radioimmunoassays were done in 900-til tubes containing 55-1il reaction mixtures. The mixture consisted of 10 ,ul of "1I antigen containing 4 ,jg of protein, 5 pl of rabbit antiserum to the antigen or nonimmune rabbit serum, and 15 Ll of ST buffer, incubated for 48 h at 5VC. The immune complex was precipitated by adding 25 p1 of sheep anti-rabbit serum (2.5 equivalents) or nonimmune sheep serum and incubating for an additional 24 h at 5VC. The precipitate was pelleted by centrifugation at 2,500 rpm in a Sorvall HL-8 rotor at 5VC for 30 min and washed twice with 500 ll of icecold 0.1 M sodium phosphate-buffered saline (pH 7.2) before counting the "I in the precipitate with a Beckman Biogamma counter. Absorption of immune sera with amA31 defective particles. Absorption of antibody from immune sera was accomplished by repeatedly adding 20 I1 of an amA31 suspension (4 mg/ml) to 300 pi of antisera, incubating for 2 h at 37°C, and centrifuging out the immune precipitate. When a pelleted precipitate was no longer detected, an additional 20 p1 of amA31 was added to the antisera and incubated at SoC for 72 h. Absorbed antisera were then centrifuged at 50,000 rpm at 5°C in a Beckman SW50 rotor to remove the excess residual virus, which appeared as a minute translucent pellet. SDS dissociation of phage. Purified phage preparations containing either 2 or 4 mg of protein per ml were dissociated in 1% SDS. Freshly dissociated phage preparations were used for each experiment. Phage was dissociated with SDS under four conditions: (i) incubation at room temperature (22°C) for 15 min before use; (iu) incubation at 37°C for 15 min; (iii) heating at 90 to 95°C for 1 min; and (iv) heating at 90 to 95°C for 4 min. Preparations were then immediately analyzed by SDS-polyacrylamide gel electrophoresis for intermediates of capsid dissociation caused by the particular treatment. SDS-polyacrylamide gel electrophoresis of intermediates of phage dissociation. Intermediates of phage dissociation in SDS were examined on SDSpolyacrylamide gels. The 8- by 0.8-cm gels containing 10% acrylamide were prepared and run according to the method of Weber and Osborn (27), except the reducing agent was usually omitted. Gels for the separation of unusually large proteins contained 2% acrylamide and 2% agarose to solidify them. Thus, for separation of phage particles from intermediates of SDS-dissociated capsids, the acrylamide concentration described for separation of nucleic acid molecules was used (15, 20) with the standard buffer system and procedure described by Weber and Osborn (27). The 50-ul samples were layered on the gels and subjected to electrophoresis at a constant current of 4.5 mA/gel.
J. VIROL.
The protein bands were fixed in 12% trichloroacetic acid and incubated for 24 h at room temperature in 20% methanol in 10% acetic acid as an additional precaution to remove possible traces of free "1I not removed by previous procedures. The gels were then cut into equal fractions, and the radioactivity was counted in a Beckman Biogamma counter.
RESULTS Purification of R17 and amA31 phage antigens. Since amA31 defective phage particles were to be used to prepare specific coat protein antibody and for the specific adsorption of coat protein antibody from R17 phage antiserum, it was essential that amA31 preparations be free of R17 phage. Normally, growth of amA31 results in production of 0.1 to 1% wild-type R17 phage revertants. When growing 20-liter batches of amA31, the number of revertants may occasionally exceed 10%. These revertants could elicit an immune response to wild-type antigenic determinants when immunizing with amA31 and consequently confuse interpretation of subsequent results. A method for elimination of wildtype phage from amA31 preparations was developed. The CsCl equilibrium centrifugation profiles of purified R17 and amA31 are shown in Fig. 1A and B. In the case of R17, a single band was seen that comigrated with the infectivity of the preparation at a density of 1.43. The amA31 preparation produced a broader absorption profile. The maximum absorption (p = 1.41) was two fractions behind the maximum infectious units and slightly skewed toward the top of the gradient. The material with a density of 1.41 has a low ratio of PFU to OD2n and represented the amA31 defective particles whose RNA is partially removed by the prior RNase treatment (8, 14, 21). Increased RNase treatment did not provide a better separation of defective amA31 particles from wild-type R17 back-revertants in the preparations. Appropriate fractions from the CsCl gradients were dialyzed and sedimented in 5 to 20% sucrose gradients. In the case of amA31, only those fractions from the CsCl gradient with a low ratio of PFU to OD2n were pooled for further processing. The sucrose gradient profiles of R17 and amA31 are shown in Fig. 2A and B. As expected, the R17 sediments as a single band with the infectivity of the preparation. The amA31 band of OD sediments slightly slower than the band of infectivity. Again, the amA31 fractions with the lowest ratio of PFU to OD2n were selected for further purification. The amA31 preparation was further purified by two cycles of filtration through a layer of E. coli S26 to remove wild-type R17 particles by
VOL. 24, 1977
A
6 0
200
To
4
100
II
-
I
2
Bottom
593
CAPSID STRUCTURE OF R17 DEFECTIVE PARTICLES
10
20
Top
Fraction Number 30
for immunization of rabbits. The recovery of defective particles from the E. coli adsorption stage in the purification ranged from 19 to 53%. A typical experiment showing the efficiency of removal of R17 infectious units from amA31 preparations at each stage during the purification is shown in Table 2. This series of steps resulted in a 100,000-fold purification of the amA31 phage preparation. The most efficient means of removal of infectious R17 from the preparation was by adsorption to E. coli. Figure 3 shows the SDS-polyacrylamide gel patterns of: (i) diphtheria toxin and toxin fragments A and B used as markers; (ii) the purified, SDS-disrupted preparation of R17; (iii) the purified, SDS-disrupted preparation of amA31 after filtration through E. coli; and (iv) the 280-nm 4
T
200
20-
20
0
2e
i10
100
Ta
200 no
k01 Iz, L4.
100 k BOttom
10
20
Top
Fraction Number
FIG. 1. Cesium chloride density gradient analysis of wild-type R17 and defective amA31 phage particles. Phage and defective particles were grown in E. coli S26. The preparations were purified as described in the text. A 1.875-g amount of CsCI was dissolved in 3 ml ofphage suspension and centrifuged at 39,000 rpm for 24 h at 100C in a Spinco SW50 rotor. The OD2m and PFU per gradient fraction were determined. (A) Wild-type R1 7; (B) defective particles of amA31. Note that infectivity does not coincide with the absorbance of defective phage in (B). The amA31 particles have a lower density than contaminant Ri 7 phage because a portion of their RNA genome has been removed by treatment with RNase.
1.5
9
!1
B
150
I-
T *
100
1.0 I!
"'Zo ZO no
Q0 czO
adsorption to the cells. The degree of purification and percent recovery of five successive preparations are shown in Table 1. Preparations 1, 2, and 3 had been partially separated from infectious R17 by CsCl and sucrose gradient centrifugation prior to the two cycles of adsorption to E. coli S26. In preparations 4 and 5, the entire OD band from a CsCl density gradient was used. Partial purification by CsCl and sucrose gradient centrifugation leads to at least a 10-fold-greater purification of amA31, as shown in Table 1. The highest degree of purity was obtained in preparation 1, where there were 1.39 x 10`0 PFU/phage particle. This preparation was used
i
R 17
50
.5
20
Bot tom
Fraction
Lu-
k
Top
Number
FIG. 2. Sedimentation profiles of wild-type R17 and defective amA31 phage particles. Fractions from cesium chloride density gradients containing wildtype R17 or amA31 fractions with low PFU-to-OD2ffi ratios were pooled, dialyzed, layered on a 28-ml 5 to 20% gradient, and centrifuged in an SW27 rotor at 27,000 rpm for 3.5 h at 5VC. The OD2f and PFU per gradient fraction were determined. (A) Wild-type R1 7; (B) defective particles of amA31.
594
IGLEWSKI
J. VIROL.
TABLE 1. Removal of wild-type Rl 7phage from preparations of defective amA31 phage particles by filtration through a layer of E. coli S26 PFU/phage particles Prepna After 1st adsorp- After 2nd adsorp- Recovery (%) tion of R17 to E. tion of R17 to E. of phagec coli coli
1.39 x 10`' 39 1.31 x 10-8 53 10-5 1.4 x 10-8 49 10-3 1.7 x 10-7 19 10-3 1.5 x 10-5 20 a Sample applied to first column ranged from 219 mg containing 4.6 x 1014 PFU to 18.3 mg containing 1011 PFU in preparation 1. Preparations 1, 2, and 3 originated from selected fractions from CsCl and sucrose gradient phage bands with low PFU/OD2w0. "One OD280 unit = 2.30 x 1013 phage particles, based on an extinction coefficient of R17 phage of 7.66 OD280/mg per ml and a virion particle weight of 3.6 x 106 daltons (5). The same extinction coefficient is assumed for amA31. e (Milligrams recovered from second column/milligrams applied to first column) x 100. 1 2 3 4 5
4.70 x 2.29 x 2.7 x 2.1 x 2.3 x
10-9
10-5
TABLE 2. Removal of wild-type R17phage from preparations of defective amA31 phage particles amA31 phage purification stage
Ratio of infectious units to physical phage particles
PFU/OD2a PFU/particlea Before CsCl gradient 1.25 x 1011 5.43 x 10-3 After CsCl gradient 4.55 x 10"' 1.98 x 10-3 After sucrose gradient 1.58 x 10'" 6.86 x 10-4 After adsorption column 8.54 x 105 3.71 x 10-8 chromatography on E. coli a One OD2s3 unit = 2.30 x 1013 phage particles, based on an extinction coefficient of R17 phage of 7.66 OD2ms/mg per ml and a virion particle weight of 3.6 x 106 daltons (5).
absorbing material at the top of the CsCl gradient shown in Fig. 1B. As expected, the R17 preparation contained two protein components. The major coat protein band (molecular weight, 13,729) was near the running front of the gel, and the minor assembly protein component (molecular weight, 38,000) comigrated with diphtheria toxin fragment B (molecular weight, 38,000). Three times the amount of amA31 as R17 protein was subjected to electrophoresis on a parallel gel. Only coat protein could be detected in amA31 preparations. In addition, no cellular proteins could be detected in amA31 preparations as a result of filtration of amA31 through the layer of E. coli. Gel 4 contains a sample from the top of the CsCl gradient shown in Fig. 1B. It contains coat protein and several highmolecular-weight compounds. This material was not used in subsequent experiments.
SDS-polyacrylamide gel electrophoresis of dissociated '25I-labeled phage preparations confirmed the results of the stained gels. They showed the assembly protein and coat protein bands from dissociated R17 phage and the complete absence of assembly protein in dissociated amA31 phage preparations (Fig. 4A and B). The dissociated amA31 phage contained a major band of coat protein that migrated in the same position as the R17 phage coat protein marker (Fig. 4A) and a minor component that migrated between coat protein and assembly protein and was found in both R17- and amA31-dissociated phage. The minor component can be detected only in preparations of labeled protein. A trace of the minor component may be found in coat protein purified from acetic acid-disrupted phage (Fig. 4C). Immunodiffusion analysis. Purified R17 phage and amA31 defective particles were tested for reactivity to R17 and amA31 antisera (Fig. 5A and B). Both antisera show a line of complete identity between two types of phage particles. The amA31 preparations show a second minor precipitin line, which shared identity with purified coat protein. The line of complete identity was more distinct after an additional 24 h of incubation. The second minor precipitin line associated with purified coat protein probably represents a polymer of coat protein subunits formed as they diffuse away from the 1% SDS dissociation reagent. Observation of coat protein polymers of MS2 during immunodiffusion have been reported (18). The diffusion of SDS from the coat protein well usually caused a partial dissociation of the phage in the adjacent wells, resulting in a broad phage precipitin line. The light haze around the coat protein well was due to the buffer in the well, since its presence was independent of antigen and antibody (data not shown). Both R17 and amA31 antisera also show a line of identity between SDS-dissociated R17 phage protein and SDS-dissociated amA31 phage protein (Fig. 6A and B). This major precipitin line shares complete identity with purified coat protein. There also appears to be a minor precipitating antibody in R17 antisera, which reacts with dissociated R17 and amA31 phage. This minor precipitin line is at the edge of the R17 antibody wells in Fig. 6A and B and is often difficult to reproduce. This minor precipitin line has never been observed in reactions between amA31 antisera and purified coat protein or dissociated R17 and amA31 phage, suggesting the wild-type and defective particles may present some different antigenic determinants to the immune system resulting in a precipitating antibody unique to R17 antisera.
g
100
q-
A
Wj.
I
50 I
I
I
I
I
10
20
3
Fraction Number
B I-,
32
L Fraction
24 r-
C
18r
;a: 4 12
-
-
Fraction Number
FIG. 3. SDS-polyacrylamide gel electrophoresis of purified R17 phage and amA31 defective particles. R17 phage and amA31 defective particles were purified as described in the text. Samples were subjected to electrophoresis at room temperature for 14 h at a constant current of 4.5 mA/gel in the buffer system described by Weber and Osborn (27). Gel 1, 100 pg of purified diphtheria toxin (molecular weight, 62,000), toxin fragment B (molecular weight, 38,000), and toxin fragment A (molecular weight, 24,000) as marker protein, provided by B. H. Iglewski; gel 2, 100 pg of SDS-dissociated R17 phage; gel 3, 300 pg of SDS-dissociated amA31 defective particles; gel 4, 300 pg of the material banding at the top of the cesium chloride density gradient shown in Fig. lB.
FIG. 4. SDS-polyacrylamide gel electrophoresis of purified antigens labeled with 12I. A 200-pg sample of each of the antigens was labeled with 1251I by the lactoperoxidase method. The samples were subjected to electrophoresis at room temperature for 6 h at a constant current of 4.5 mA/gel. (A) 100 pg of SDSdissociated Rl7phage; (B) 100 pg of SDS-dissociated amA31-defective particles; and (C) 50 pl ofpurified coat protein prepared from 12'I-labeled R17 phage according to the method of Sugiyama et al. (22, 23). Since purified coat protein is unstable in solution and a protein determination was not done before running these gels, the precise amount of coat protein in the gel sample is unknown. Its original protein concentration was 170 pg/100 p1. Nevertheless, purified coat protein serves as an adequate marker for these gels.
595
596
IGLEWSKI
J. VIROL.
FIG. 5. Antigenic relationship between R17 phage and amA31 defective particles. Gel immunodiffusion slides without SDS were prepared as described in the text. The wells were filled with 10 4d of the various preparations. (A) Center well: R17 antisera. Outer wells beginning at top and going clockwise (milligrams per milliliter): R1 7, 2; amA31, 2; R1 7, 2; amA31, 2; R1 7, 2; coat protein in SDS-/3-mercaptoethanol-containing buffer, 1.5. (B) Center well: amA31 antisera. Outer well beginning at top (milligrams per milliliter): R17, 2; coat protein in SDS-fi-mercaptoethanol-containing buffer, 1.5; amA31, 2; R17, 2; amA31, 2; R17, 2.
FIG. 6. Antigenic relationship between SDS-dissociated R17 and amA31 defective particles. Gel immunodiffusion slides contained 0.1% SDS as described in the text. Antigens were dissociated in 1% SDS at 90'C for I min at an antigen concentration of 1 or 2 mg/ml. An antigen concentration of 1 mg/ml is shown because little or no nonspecific precipitation occurred around the outer antibody wells under these conditions. Wells were filled with 10 4l of the reagents. (A) Center well: SDS-dissociated R17phage. Contents of outer wells, beginning at top and going clockwise: R17 antiserum; amA31 antiserum; R17 antiserum; empty; nonimmune serum; purified coat protein. (B) Center well: SDS-dissociated amA31-defective particles. Contents of outer wells beginning at top: amA31 antiserum; purified coat protein; nonimmune serum; empty; amA31 antiserum; R1 7 antiserum.
SDS, at concentrations used for polyacrylamide gel electrophoresis, was incorporated into immunodiffusion agar to insure solubilization and diffusion of the spontaneously precipitable, dissociated phage protein. The SDS did not interfere with the coat protein precipitin reaction but did have the disadvantage of causing a slight precipitate to form around some antisera wells. The intensity of the precipitate varied between serum samples and may reflect components in the rabbits' diet before bleeding or serum K+ and Ca2" levels. Potassium and cal-
cium salts of SDS are insoluble. Dialysis of the antisera against sodium phosphate-buffered saline before use reduced the intensity of the precipitate but did not always completely eliminate it. Rabbits immunized with assembly protein, purified from disrupted phage according to the method of Steitz (21), did not produce antibody detectable in immunodiffusion assays. The antisera were absorbed with amA31 before assem-
bly protein antibody was assayed. Radioimmunoassay. Primary antibody-an-
VOL. 24, 1977
CAPSID STRUCTURE OF R17 DEFECTIVE PARTICLES
were studied by radioimmuAntisera to R17 and purified assembly protein were adsorbed with amA31 defective particles. The precipitated immune complex and excess amA31 were removed by centrifugation. Both absorbed and nonabsorbed antisera were incubated with either "2I-labeled R17 or amA31, and then sheep anti-rabbit serum was added. An R17 antibody was detected in antiserum that had been absorbed with an excess of amA31 (Table 3). The antibody remaining in the absorbed serum bound to R17 but did not bind to amA31, demonstrating the presence of antigenic determinants on R17 that are not available for binding antibody on amA31. In contrast, antiserum from rabbits immunized with purified assembly protein contained sufficient antibody to precipitate 56 to 60% of labeled R17 or amA31. This antibody could be completely absorbed with amA31 and most likely represents antibody to coat protein contaminants in the assembly protein antigen. Neutralization of R17 phage. The effect of the various antisera on phage neutralization was examined (Table 4). Both R17 and amA31 antisera completely neutralized the infectivity of R17 phage during a 30-min incubation at 37°C. The R17 antiserum that was previously absorbed with amA31 had no neutralizing activity. Although antisera from rabbits immunized with purified assembly protein neutralized R17 phage, this activity could also be eliminated by
tigen interactions
TABLE 4. Neutralization of Rl 7phagea
noassay.
TABLE 3. Radioimmunoassay for R17 antibody not absorbed by amA31 % Total antigen precipitated
Antisera to:
R17b
amA31c
R17 capsids 99 98 R17 capsids absorbed with R17 0 NDd R17 capsids absorbed with amA31 35 2 Purified assembly proteine 56 60 Purified assembly protein absorbed 0 0 with amA31 a The percentage of total counts precipitated is the average of triplicate samples from which background radioactivity has been subtracted. The reactions contained l"I-labeled antigen + rabbit antibody + sheep anti-rabbit serum. Background was determined by incubating labeled antigen with sheep anti-rabbit serum as well as incubating labeled antigen with normal rabbit serum and sheep anti-rabbit serum. b "2MI-abeled R17 containing 7 x 105 cpm/4 fyg per reaction. c 125I-labeled amA31 containing 2 x 105 cpm/4 ug per d
reaction. ND, Not done.
e Assembly protein preparations with traces of coat protein (10).
are
contaminated
597
Antiserum
Titer
Normal rabbit serum ................ 4.5 x 102 R17 phage antiserum
JOURNAL OF VIROLOGY, Nov. 1977, p. 590-601 Copyright © 1977 American Society for Microbiology
Changes in the Capsid Structure and Stability of Defective Particles of Bacteriophage R17 WALLACE J. IGLEWSKI Department of Microbiology and Immunology, University of Oregon Health Sciences Center, Portland, Oregon 97201 Received for publication 28 April 1977
Serological and chemical methods were used to compare the capsid structure and stability of R17 phage and amA31 defective particles. Immunodiffusion analysis demonstrated identity between intact R17 and amA31 capsids and between dissociated subunits of both R17 and amA31 and purified coat protein. Radioimmunoassays detected an antibody in R17 antisera that binds to intact R17 but could not be absorbed from R17 antisera with amA31. The R17 antibody remaining in amA31-absorbed sera did not neutralize infectivity of R17 phage. Differences between the surface composition of R17 and amA31 capsids were also detected by iodination. Capsids of R17 bound approximately four times more 125I than amA31, which was accounted for by a decreased 125I labeling of coat protein. Finally, amA31 capsids dissociated under milder conditions of sodium dodecyl sulfate treatment than R17 capsids. The sodium dodecyl sulfate dissociation of both R17 and amA31 capsids resulted in the formation of a transient 38,000-dalton intermediate, which subsequently dissociated to coat protein monomers. Preparations of dissociated R17 capsids also contained assembly protein and a protein of approximately 27,000 daltons. The 27,000-dalton protein was also found in preparations of dissociated amA31 capsids.
Bacteriophage R17 is a small icosahedral phage having a diameter of about 13.5 nm (9). The phage has a single-stranded RNA genome, which is surrounded by 180 coat protein subunits and 1 assembly protein subunit for attachment to susceptible cells (7, 14, 21, 25). Recently, it has been shown that assembly protein is located on the surface of the phage (2), which is compatible with its role as an attachment site. In terms of the integrity of the phage, assembly protein participates in maintaining the barrier between the phage genome and the external environment, since the genome of mutants that lack assembly protein is degraded by treatment with RNase (8, 14). In this study the organization and stability of the coat protein capsid of the R17 phage was compared with an amber mutant that lacks assembly protein. The absence of assembly protein alters the surface of the capsid and decreases the stability of the capsid as measured by dissociation with sodium dodecyl sulfate (SDS). Thus, assembly protein appears to also have a general effect on the capsid structure. MATERIALS AND METHODS
richia coli S26 and S26RIE, respectively, were provided by R. Kamen. The viral mutant was grown and plaqued on E. coli S26RIE by the method of Tooze and Weber (24) and contained less than 1.0% revertants. E. coli S26 was used to grow wild-type R17 and produce defective particles of amA31 lacking the assembly protein after infection with complete amA31 grown in the permissive cells. Growth and purification of wild-type R17 and defective amA31 phage. Growth of large batches of R17 and its purification are based on the published procedure of Vasquez et al. (25). Briefly, batches (20liter) of E. coli S26 grown in MS broth (20) to 3 x 108 cells per ml were infected with 10 PFU of R17 per cell. After 3 h of incubation, the cells were chilled and lysed by shaking with 10% (vol/vol) chloroform. The lysate was decanted, and 350 g per liter of ammonium sulfate was added. The precipitate was harvested by continuous-flow centrifugation in a Sorvall SS34 rotor at 18,000 rpm. The pellet was resuspended in 300 ml of 0.1 M NaCl and 0.05 M Tris, pH 7.2 (ST), buffer, and the phage was further purified by two cycles of differential centrifugation at 10,000 rpm in a Sorvall SS34 rotor for 20 min to remove debris and at 50,000 rpm in a Beckman SW50 rotor for 90 min to sediment the phage. The phage was resuspended in ST buffer containing 10-3 M MgCl2 and treated with a mixture of 5,ug of RNase A and 5 ,ug of DNase per ml for 30 min at 37°C. This was followed by digestion with 25 ,g of Pronase per ml at 25°C for 30 min. The phage was sedimented as described above and resuspended in ST buffer for CsCl equilibrium
Bacteria and bacteriophage. Amber mutant amA31 of the bacteriophage R17, along with the nonpermissive and permissive indicator strains, Esche590
VOL. 24, 1977
CAPSID STRUCTURE OF R17 DEFECTIVE PARTICLES
centrifugation at 39,000 rpm in a Beckman SW50 rotor for 24 h. The virus that banded at p = 1.43 was collected and dialyzed against ST buffer before being centrifuged through a 5 to 20% sucrose gradient containing ST buffer (11) at 27,000 rpm for 3.5 h at 5VC in a Beckman SW27 rotor. Phage bands in both CsCl and sucrose gradients were monitored by optical density at 280 nm (OD20) and PFU per fraction. The buoyant density of the phage was determined by measuring the CsCl concentration of fractions with a refractometer. Defective amA31 phage particles were produced in E. coli S26, the nonpermissive host, and purified as described above. In the CsCl gradient, the defective particles banded at a buoyant density slightly lower than infectious units due to a partial loss of their RNA genome during treatment with RNase A (21). Likewise, the defective particles sedimented slightly slower than the back-revertant, wild-type phage in a sucrose gradient (9). Selection of defective-particle fractions with minimum infectivity per unit of adsorbance at 280 nm within the phage band resulted in partial purification of defective particles from intact phage. The final step in purification of the amA31 defective particles from wild-type phage made use of the inability of assembly protein-deficient mutants to adsorb to E. coli S26 (14). Cells were grown to 4 x 108 cells per ml in MS and harvested by centrifugation at 10,000 rpm for 10 min in a Sorvall SS34 rotor. The cells were washed and resuspended in physiological saline containing 5 ml of 1 M CaCl2 and 2 ml of 50% glucose per liter. The cells were collected on a 142mm membrane filter (Millipore Corp.), 0.22-,um pore size. Eight sheets of Whatman no. 4 filter paper were used as a prefilter to prevent clogging of the Millipore filter. Glucose and CaCl2 were added to the phage suspension at the concentrations indicated above, warmed at 37°C, and slowly filtered through the layer of cells. The defective particles passed through the filter, whereas the revertant, wild-type R17 adsorbed to the cells as expected (14). The cell layer was washed with physiological saline containing CaCl2 and glucose to flush through residual defective phage. Purity of amA31 preparations was determined by the decreased number of PFU per OD280 of the phage suspension and SDS-polyacrylamide gel electrophoresis of nIlabeled protein from SDS-disrupted phage. Lactoperoxidase-catalyzed iodination of R17 and amA31 defective phage. Enzymatic iodination of purified phage particles was done essentially as described by Gibson (6). To 100,ul of purified phage containing 100 Mg of protein, the following sequential additions were made: (i) 10 IlI of carrier-free '5I; (ii) 20 M1 of lactoperoxidase bound to Sepharose beads (approximately 10,ug of bound lactoperoxidase); (iii) 10Ml of 3.3 x 10-5 M KI and 10 ,ul of a 1:2,000 dilution of 30% H202 in distilled water. The reaction was allowed to proceed for 30 min at 22°C with gentle shaking to keep the Sepharose beads suspended. Under these conditions the reaction should be essentially completed in 10 to 15 min (3). The Sepharose beadbound lactoperoxidase was removed by centrifugation at 500 x g for 10 min at 5°C. Free "1I was removed during subsequent rate-zonal sucrose gradient sedimentation of the phage as described above. The phage
591
band was diluted in ST buffer and then pelleted by centrifugation at 50,000 rpm for 90 min in a Beckman SW50 rotor. Preparation of R17 coat protein. The cold acetic acid method of Sugiyama et al. (22, 23) was used to isolate coat protein from purified R17 phage. One volume of ice-cold R17 phage (6 mg/ml) in ST buffer was added to 2 volumes of cold glacial acetic acid and incubated in an ice bath for 1 h. The precipitated RNA and assembly protein were then removed by centrifugation at 8,000 rpm in a Sorvall SS34 rotor for 20 min at 5°C. The protein solution was dialyzed against three to four changes of 0.001 M acetic acid at 5°C until the pH of the dialysate reached 3.2 to 3.5. The coat protein solution was centrifuged at 50,000 rpm in an SW50 rotor at 5°C for 90 min to remove possible traces of phage particles that escaped the acid treatment and then dialyzed extensively against a buffer containing 0.01 M sodium phosphate (pH 7.2), 0.29 M /?-mercaptoethanol, and 0.1% SDS. The coat protein was stable for several weeks at room temperature when stored under nitrogen in a sealed tube. Preparations of l"I-labeled coat protein were obtained from previously iodinated phage. SDS-polyacrylamide gel electrophoresis. Purity of phage preparations was examined by analysis on SDS-polyacrylamide gels. The 17- by 0.8-cm or 8by 0.8-cm gels were prepared and run according to the method of Weber and Osborn (27), except the reducing agent was omitted when phage protein was analyzed. Phage samples were dissociated by heating for 1 min at 90°C in 1.0% SDS. The protein bands in the gels were fixed and stained in 0.7% Coomassie brilliant blue, and the gel was destined by diffusion in a solution of 20% methanol, 10% acetic acid, and 70% water. Proteins labeled with "1I were detected by counting gel fractions in a Beckman Biogamma counter. Purification of assembly protein. Assembly protein was purified according to the procedure of Steitz (21). The procedure consists of solubilizing purified phage in 6 M guanidine-HCl, dialyzing into 8 M urea, and eluting from a cellulose phosphate column with a linear NaCl gradient. Assembly protein elutes in 8 M urea-0.05 M Tris-hydrochloride (pH 8.5) containing 0.2 M NaCl. The protein must be kept in high concentrations of urea, guanidine-HCl, or SDS to prevent precipitation. Preparation of antibodies. Antisera to R17 and amA31 were made in 2-kg New Zealand white rabbits by injecting 0.1 mg of phage in complete Freund adjuvant into four subcutaneous sites on the back and in two rear footpads. The rabbits were immunized in the same manner 3 weeks later with phage in incomplete Freund adjuvant and bled after 10 days. Antiserum to assembly protein was made by emulsifying the assembly protein fractions (0.1 mg of protein) from the cellulose phosphate column with an equal volume of complete Freund adjuvant and inoculating and bleeding rabbits on the same schedule as described above. Immunodiffusion analysis. Immunodiffusion (15) was done at room temperature on glass microscope slides using 0.75% agarose in 0.01 M phosphate
592
IGLEWSKI
buffer (pH 7.5) containing 0.5 M glycine, 0.14 M sodium citrate, and 0.02% Merthiolate. Approximately 2.5 ml of the agarose solution (sometimes containing 0.1% SDS) was layered onto each slide and allowed to harden, and wells were cut out of the agar. Incorporation of SDS in diffusion agar helped prevent aggregation and precipitation of coat protein antigen when SDS-dissociated phage antigens were used (28). The slides were placed in a moist chamber to retard drying. Radioimmunoassay for phage antibody. Radioimmunoassays were done in 900-til tubes containing 55-1il reaction mixtures. The mixture consisted of 10 ,ul of "1I antigen containing 4 ,jg of protein, 5 pl of rabbit antiserum to the antigen or nonimmune rabbit serum, and 15 Ll of ST buffer, incubated for 48 h at 5VC. The immune complex was precipitated by adding 25 p1 of sheep anti-rabbit serum (2.5 equivalents) or nonimmune sheep serum and incubating for an additional 24 h at 5VC. The precipitate was pelleted by centrifugation at 2,500 rpm in a Sorvall HL-8 rotor at 5VC for 30 min and washed twice with 500 ll of icecold 0.1 M sodium phosphate-buffered saline (pH 7.2) before counting the "I in the precipitate with a Beckman Biogamma counter. Absorption of immune sera with amA31 defective particles. Absorption of antibody from immune sera was accomplished by repeatedly adding 20 I1 of an amA31 suspension (4 mg/ml) to 300 pi of antisera, incubating for 2 h at 37°C, and centrifuging out the immune precipitate. When a pelleted precipitate was no longer detected, an additional 20 p1 of amA31 was added to the antisera and incubated at SoC for 72 h. Absorbed antisera were then centrifuged at 50,000 rpm at 5°C in a Beckman SW50 rotor to remove the excess residual virus, which appeared as a minute translucent pellet. SDS dissociation of phage. Purified phage preparations containing either 2 or 4 mg of protein per ml were dissociated in 1% SDS. Freshly dissociated phage preparations were used for each experiment. Phage was dissociated with SDS under four conditions: (i) incubation at room temperature (22°C) for 15 min before use; (iu) incubation at 37°C for 15 min; (iii) heating at 90 to 95°C for 1 min; and (iv) heating at 90 to 95°C for 4 min. Preparations were then immediately analyzed by SDS-polyacrylamide gel electrophoresis for intermediates of capsid dissociation caused by the particular treatment. SDS-polyacrylamide gel electrophoresis of intermediates of phage dissociation. Intermediates of phage dissociation in SDS were examined on SDSpolyacrylamide gels. The 8- by 0.8-cm gels containing 10% acrylamide were prepared and run according to the method of Weber and Osborn (27), except the reducing agent was usually omitted. Gels for the separation of unusually large proteins contained 2% acrylamide and 2% agarose to solidify them. Thus, for separation of phage particles from intermediates of SDS-dissociated capsids, the acrylamide concentration described for separation of nucleic acid molecules was used (15, 20) with the standard buffer system and procedure described by Weber and Osborn (27). The 50-ul samples were layered on the gels and subjected to electrophoresis at a constant current of 4.5 mA/gel.
J. VIROL.
The protein bands were fixed in 12% trichloroacetic acid and incubated for 24 h at room temperature in 20% methanol in 10% acetic acid as an additional precaution to remove possible traces of free "1I not removed by previous procedures. The gels were then cut into equal fractions, and the radioactivity was counted in a Beckman Biogamma counter.
RESULTS Purification of R17 and amA31 phage antigens. Since amA31 defective phage particles were to be used to prepare specific coat protein antibody and for the specific adsorption of coat protein antibody from R17 phage antiserum, it was essential that amA31 preparations be free of R17 phage. Normally, growth of amA31 results in production of 0.1 to 1% wild-type R17 phage revertants. When growing 20-liter batches of amA31, the number of revertants may occasionally exceed 10%. These revertants could elicit an immune response to wild-type antigenic determinants when immunizing with amA31 and consequently confuse interpretation of subsequent results. A method for elimination of wildtype phage from amA31 preparations was developed. The CsCl equilibrium centrifugation profiles of purified R17 and amA31 are shown in Fig. 1A and B. In the case of R17, a single band was seen that comigrated with the infectivity of the preparation at a density of 1.43. The amA31 preparation produced a broader absorption profile. The maximum absorption (p = 1.41) was two fractions behind the maximum infectious units and slightly skewed toward the top of the gradient. The material with a density of 1.41 has a low ratio of PFU to OD2n and represented the amA31 defective particles whose RNA is partially removed by the prior RNase treatment (8, 14, 21). Increased RNase treatment did not provide a better separation of defective amA31 particles from wild-type R17 back-revertants in the preparations. Appropriate fractions from the CsCl gradients were dialyzed and sedimented in 5 to 20% sucrose gradients. In the case of amA31, only those fractions from the CsCl gradient with a low ratio of PFU to OD2n were pooled for further processing. The sucrose gradient profiles of R17 and amA31 are shown in Fig. 2A and B. As expected, the R17 sediments as a single band with the infectivity of the preparation. The amA31 band of OD sediments slightly slower than the band of infectivity. Again, the amA31 fractions with the lowest ratio of PFU to OD2n were selected for further purification. The amA31 preparation was further purified by two cycles of filtration through a layer of E. coli S26 to remove wild-type R17 particles by
VOL. 24, 1977
A
6 0
200
To
4
100
II
-
I
2
Bottom
593
CAPSID STRUCTURE OF R17 DEFECTIVE PARTICLES
10
20
Top
Fraction Number 30
for immunization of rabbits. The recovery of defective particles from the E. coli adsorption stage in the purification ranged from 19 to 53%. A typical experiment showing the efficiency of removal of R17 infectious units from amA31 preparations at each stage during the purification is shown in Table 2. This series of steps resulted in a 100,000-fold purification of the amA31 phage preparation. The most efficient means of removal of infectious R17 from the preparation was by adsorption to E. coli. Figure 3 shows the SDS-polyacrylamide gel patterns of: (i) diphtheria toxin and toxin fragments A and B used as markers; (ii) the purified, SDS-disrupted preparation of R17; (iii) the purified, SDS-disrupted preparation of amA31 after filtration through E. coli; and (iv) the 280-nm 4
T
200
20-
20
0
2e
i10
100
Ta
200 no
k01 Iz, L4.
100 k BOttom
10
20
Top
Fraction Number
FIG. 1. Cesium chloride density gradient analysis of wild-type R17 and defective amA31 phage particles. Phage and defective particles were grown in E. coli S26. The preparations were purified as described in the text. A 1.875-g amount of CsCI was dissolved in 3 ml ofphage suspension and centrifuged at 39,000 rpm for 24 h at 100C in a Spinco SW50 rotor. The OD2m and PFU per gradient fraction were determined. (A) Wild-type R1 7; (B) defective particles of amA31. Note that infectivity does not coincide with the absorbance of defective phage in (B). The amA31 particles have a lower density than contaminant Ri 7 phage because a portion of their RNA genome has been removed by treatment with RNase.
1.5
9
!1
B
150
I-
T *
100
1.0 I!
"'Zo ZO no
Q0 czO
adsorption to the cells. The degree of purification and percent recovery of five successive preparations are shown in Table 1. Preparations 1, 2, and 3 had been partially separated from infectious R17 by CsCl and sucrose gradient centrifugation prior to the two cycles of adsorption to E. coli S26. In preparations 4 and 5, the entire OD band from a CsCl density gradient was used. Partial purification by CsCl and sucrose gradient centrifugation leads to at least a 10-fold-greater purification of amA31, as shown in Table 1. The highest degree of purity was obtained in preparation 1, where there were 1.39 x 10`0 PFU/phage particle. This preparation was used
i
R 17
50
.5
20
Bot tom
Fraction
Lu-
k
Top
Number
FIG. 2. Sedimentation profiles of wild-type R17 and defective amA31 phage particles. Fractions from cesium chloride density gradients containing wildtype R17 or amA31 fractions with low PFU-to-OD2ffi ratios were pooled, dialyzed, layered on a 28-ml 5 to 20% gradient, and centrifuged in an SW27 rotor at 27,000 rpm for 3.5 h at 5VC. The OD2f and PFU per gradient fraction were determined. (A) Wild-type R1 7; (B) defective particles of amA31.
594
IGLEWSKI
J. VIROL.
TABLE 1. Removal of wild-type Rl 7phage from preparations of defective amA31 phage particles by filtration through a layer of E. coli S26 PFU/phage particles Prepna After 1st adsorp- After 2nd adsorp- Recovery (%) tion of R17 to E. tion of R17 to E. of phagec coli coli
1.39 x 10`' 39 1.31 x 10-8 53 10-5 1.4 x 10-8 49 10-3 1.7 x 10-7 19 10-3 1.5 x 10-5 20 a Sample applied to first column ranged from 219 mg containing 4.6 x 1014 PFU to 18.3 mg containing 1011 PFU in preparation 1. Preparations 1, 2, and 3 originated from selected fractions from CsCl and sucrose gradient phage bands with low PFU/OD2w0. "One OD280 unit = 2.30 x 1013 phage particles, based on an extinction coefficient of R17 phage of 7.66 OD280/mg per ml and a virion particle weight of 3.6 x 106 daltons (5). The same extinction coefficient is assumed for amA31. e (Milligrams recovered from second column/milligrams applied to first column) x 100. 1 2 3 4 5
4.70 x 2.29 x 2.7 x 2.1 x 2.3 x
10-9
10-5
TABLE 2. Removal of wild-type R17phage from preparations of defective amA31 phage particles amA31 phage purification stage
Ratio of infectious units to physical phage particles
PFU/OD2a PFU/particlea Before CsCl gradient 1.25 x 1011 5.43 x 10-3 After CsCl gradient 4.55 x 10"' 1.98 x 10-3 After sucrose gradient 1.58 x 10'" 6.86 x 10-4 After adsorption column 8.54 x 105 3.71 x 10-8 chromatography on E. coli a One OD2s3 unit = 2.30 x 1013 phage particles, based on an extinction coefficient of R17 phage of 7.66 OD2ms/mg per ml and a virion particle weight of 3.6 x 106 daltons (5).
absorbing material at the top of the CsCl gradient shown in Fig. 1B. As expected, the R17 preparation contained two protein components. The major coat protein band (molecular weight, 13,729) was near the running front of the gel, and the minor assembly protein component (molecular weight, 38,000) comigrated with diphtheria toxin fragment B (molecular weight, 38,000). Three times the amount of amA31 as R17 protein was subjected to electrophoresis on a parallel gel. Only coat protein could be detected in amA31 preparations. In addition, no cellular proteins could be detected in amA31 preparations as a result of filtration of amA31 through the layer of E. coli. Gel 4 contains a sample from the top of the CsCl gradient shown in Fig. 1B. It contains coat protein and several highmolecular-weight compounds. This material was not used in subsequent experiments.
SDS-polyacrylamide gel electrophoresis of dissociated '25I-labeled phage preparations confirmed the results of the stained gels. They showed the assembly protein and coat protein bands from dissociated R17 phage and the complete absence of assembly protein in dissociated amA31 phage preparations (Fig. 4A and B). The dissociated amA31 phage contained a major band of coat protein that migrated in the same position as the R17 phage coat protein marker (Fig. 4A) and a minor component that migrated between coat protein and assembly protein and was found in both R17- and amA31-dissociated phage. The minor component can be detected only in preparations of labeled protein. A trace of the minor component may be found in coat protein purified from acetic acid-disrupted phage (Fig. 4C). Immunodiffusion analysis. Purified R17 phage and amA31 defective particles were tested for reactivity to R17 and amA31 antisera (Fig. 5A and B). Both antisera show a line of complete identity between two types of phage particles. The amA31 preparations show a second minor precipitin line, which shared identity with purified coat protein. The line of complete identity was more distinct after an additional 24 h of incubation. The second minor precipitin line associated with purified coat protein probably represents a polymer of coat protein subunits formed as they diffuse away from the 1% SDS dissociation reagent. Observation of coat protein polymers of MS2 during immunodiffusion have been reported (18). The diffusion of SDS from the coat protein well usually caused a partial dissociation of the phage in the adjacent wells, resulting in a broad phage precipitin line. The light haze around the coat protein well was due to the buffer in the well, since its presence was independent of antigen and antibody (data not shown). Both R17 and amA31 antisera also show a line of identity between SDS-dissociated R17 phage protein and SDS-dissociated amA31 phage protein (Fig. 6A and B). This major precipitin line shares complete identity with purified coat protein. There also appears to be a minor precipitating antibody in R17 antisera, which reacts with dissociated R17 and amA31 phage. This minor precipitin line is at the edge of the R17 antibody wells in Fig. 6A and B and is often difficult to reproduce. This minor precipitin line has never been observed in reactions between amA31 antisera and purified coat protein or dissociated R17 and amA31 phage, suggesting the wild-type and defective particles may present some different antigenic determinants to the immune system resulting in a precipitating antibody unique to R17 antisera.
g
100
q-
A
Wj.
I
50 I
I
I
I
I
10
20
3
Fraction Number
B I-,
32
L Fraction
24 r-
C
18r
;a: 4 12
-
-
Fraction Number
FIG. 3. SDS-polyacrylamide gel electrophoresis of purified R17 phage and amA31 defective particles. R17 phage and amA31 defective particles were purified as described in the text. Samples were subjected to electrophoresis at room temperature for 14 h at a constant current of 4.5 mA/gel in the buffer system described by Weber and Osborn (27). Gel 1, 100 pg of purified diphtheria toxin (molecular weight, 62,000), toxin fragment B (molecular weight, 38,000), and toxin fragment A (molecular weight, 24,000) as marker protein, provided by B. H. Iglewski; gel 2, 100 pg of SDS-dissociated R17 phage; gel 3, 300 pg of SDS-dissociated amA31 defective particles; gel 4, 300 pg of the material banding at the top of the cesium chloride density gradient shown in Fig. lB.
FIG. 4. SDS-polyacrylamide gel electrophoresis of purified antigens labeled with 12I. A 200-pg sample of each of the antigens was labeled with 1251I by the lactoperoxidase method. The samples were subjected to electrophoresis at room temperature for 6 h at a constant current of 4.5 mA/gel. (A) 100 pg of SDSdissociated Rl7phage; (B) 100 pg of SDS-dissociated amA31-defective particles; and (C) 50 pl ofpurified coat protein prepared from 12'I-labeled R17 phage according to the method of Sugiyama et al. (22, 23). Since purified coat protein is unstable in solution and a protein determination was not done before running these gels, the precise amount of coat protein in the gel sample is unknown. Its original protein concentration was 170 pg/100 p1. Nevertheless, purified coat protein serves as an adequate marker for these gels.
595
596
IGLEWSKI
J. VIROL.
FIG. 5. Antigenic relationship between R17 phage and amA31 defective particles. Gel immunodiffusion slides without SDS were prepared as described in the text. The wells were filled with 10 4d of the various preparations. (A) Center well: R17 antisera. Outer wells beginning at top and going clockwise (milligrams per milliliter): R1 7, 2; amA31, 2; R1 7, 2; amA31, 2; R1 7, 2; coat protein in SDS-/3-mercaptoethanol-containing buffer, 1.5. (B) Center well: amA31 antisera. Outer well beginning at top (milligrams per milliliter): R17, 2; coat protein in SDS-fi-mercaptoethanol-containing buffer, 1.5; amA31, 2; R17, 2; amA31, 2; R17, 2.
FIG. 6. Antigenic relationship between SDS-dissociated R17 and amA31 defective particles. Gel immunodiffusion slides contained 0.1% SDS as described in the text. Antigens were dissociated in 1% SDS at 90'C for I min at an antigen concentration of 1 or 2 mg/ml. An antigen concentration of 1 mg/ml is shown because little or no nonspecific precipitation occurred around the outer antibody wells under these conditions. Wells were filled with 10 4l of the reagents. (A) Center well: SDS-dissociated R17phage. Contents of outer wells, beginning at top and going clockwise: R17 antiserum; amA31 antiserum; R17 antiserum; empty; nonimmune serum; purified coat protein. (B) Center well: SDS-dissociated amA31-defective particles. Contents of outer wells beginning at top: amA31 antiserum; purified coat protein; nonimmune serum; empty; amA31 antiserum; R1 7 antiserum.
SDS, at concentrations used for polyacrylamide gel electrophoresis, was incorporated into immunodiffusion agar to insure solubilization and diffusion of the spontaneously precipitable, dissociated phage protein. The SDS did not interfere with the coat protein precipitin reaction but did have the disadvantage of causing a slight precipitate to form around some antisera wells. The intensity of the precipitate varied between serum samples and may reflect components in the rabbits' diet before bleeding or serum K+ and Ca2" levels. Potassium and cal-
cium salts of SDS are insoluble. Dialysis of the antisera against sodium phosphate-buffered saline before use reduced the intensity of the precipitate but did not always completely eliminate it. Rabbits immunized with assembly protein, purified from disrupted phage according to the method of Steitz (21), did not produce antibody detectable in immunodiffusion assays. The antisera were absorbed with amA31 before assem-
bly protein antibody was assayed. Radioimmunoassay. Primary antibody-an-
VOL. 24, 1977
CAPSID STRUCTURE OF R17 DEFECTIVE PARTICLES
were studied by radioimmuAntisera to R17 and purified assembly protein were adsorbed with amA31 defective particles. The precipitated immune complex and excess amA31 were removed by centrifugation. Both absorbed and nonabsorbed antisera were incubated with either "2I-labeled R17 or amA31, and then sheep anti-rabbit serum was added. An R17 antibody was detected in antiserum that had been absorbed with an excess of amA31 (Table 3). The antibody remaining in the absorbed serum bound to R17 but did not bind to amA31, demonstrating the presence of antigenic determinants on R17 that are not available for binding antibody on amA31. In contrast, antiserum from rabbits immunized with purified assembly protein contained sufficient antibody to precipitate 56 to 60% of labeled R17 or amA31. This antibody could be completely absorbed with amA31 and most likely represents antibody to coat protein contaminants in the assembly protein antigen. Neutralization of R17 phage. The effect of the various antisera on phage neutralization was examined (Table 4). Both R17 and amA31 antisera completely neutralized the infectivity of R17 phage during a 30-min incubation at 37°C. The R17 antiserum that was previously absorbed with amA31 had no neutralizing activity. Although antisera from rabbits immunized with purified assembly protein neutralized R17 phage, this activity could also be eliminated by
tigen interactions
TABLE 4. Neutralization of Rl 7phagea
noassay.
TABLE 3. Radioimmunoassay for R17 antibody not absorbed by amA31 % Total antigen precipitated
Antisera to:
R17b
amA31c
R17 capsids 99 98 R17 capsids absorbed with R17 0 NDd R17 capsids absorbed with amA31 35 2 Purified assembly proteine 56 60 Purified assembly protein absorbed 0 0 with amA31 a The percentage of total counts precipitated is the average of triplicate samples from which background radioactivity has been subtracted. The reactions contained l"I-labeled antigen + rabbit antibody + sheep anti-rabbit serum. Background was determined by incubating labeled antigen with sheep anti-rabbit serum as well as incubating labeled antigen with normal rabbit serum and sheep anti-rabbit serum. b "2MI-abeled R17 containing 7 x 105 cpm/4 fyg per reaction. c 125I-labeled amA31 containing 2 x 105 cpm/4 ug per d
reaction. ND, Not done.
e Assembly protein preparations with traces of coat protein (10).
are
contaminated
597
Antiserum
Titer
Normal rabbit serum ................ 4.5 x 102 R17 phage antiserum
